Open-ended VQA benchmarking of Vision-Language models by exploiting Classification datasets and their semantic hierarchy
This work addresses the problem of precise benchmarking for vision-language models, which is incremental as it builds on existing classification datasets and evaluation methods.
The paper tackled the challenge of evaluating text-generative vision-language models by proposing a novel VQA benchmark based on classification datasets, using semantic hierarchies for follow-up questions and comparing metrics, and applied it to show detailed comparisons of model abilities on object, action, and attribute classification.
The evaluation of text-generative vision-language models is a challenging yet crucial endeavor. By addressing the limitations of existing Visual Question Answering (VQA) benchmarks and proposing innovative evaluation methodologies, our research seeks to advance our understanding of these models' capabilities. We propose a novel VQA benchmark based on well-known visual classification datasets which allows a granular evaluation of text-generative vision-language models and their comparison with discriminative vision-language models. To improve the assessment of coarse answers on fine-grained classification tasks, we suggest using the semantic hierarchy of the label space to ask automatically generated follow-up questions about the ground-truth category. Finally, we compare traditional NLP and LLM-based metrics for the problem of evaluating model predictions given ground-truth answers. We perform a human evaluation study upon which we base our decision on the final metric. We apply our benchmark to a suite of vision-language models and show a detailed comparison of their abilities on object, action, and attribute classification. Our contributions aim to lay the foundation for more precise and meaningful assessments, facilitating targeted progress in the exciting field of vision-language modeling.